qiVqdZddlZddlZddlZddlZddlmZmZddlm Z m Z m Z m Z ddl Z ddlmZddlmZddlmZmZmZmZmZmZmZmZmZddlmZmZm Z m!Z!m"Z"dd l#m$Z$m%Z%m&Z&m'Z'dd l(m)Z)dd l*m+Z+dd l,m-Z-m.Z.dd l/m0Z0ddl1m2Z2 ddl3Z3ddl4m5Z6ddl7m8Z8de-de e=fdZ>GddZ?eGdde%Z@Gdde+ZAy#e9$r7Z:ejvde:ejvdedtj@d tjBd!tjDd"itjFd#tjHd$tjJd%tjLd&tjNd'tjPd(tjRd)tjTd*tjVd+tjXd,tjZd-tj\d.tj^d/tj`d0tjbd1tjdd2tjfd3itjhd4tjjd5tjld6tjnd7tjpd8tjrd9tjtd:tjvd;tjxd<tjzd=tj|d>tj~d?tjd@tjdAtjdBtjdCtjdDitjdEtjdFtjdGtjdHtjdItjdJtjdKtjdLtjdMtjdNtjdOtjdPtjdQtjdRtjdStjdTtjdUtjdVtjdWtjdXtjdYtjdZtjd[tjd\tjd]tjd^tjd_tjd`tjdatjdbtjdci}t||ddeS)fzConvert a Language enum to Gladia's language code format. Args: language: The Language enum value to convert. Returns: The Gladia language code string or None if not supported. afamarasazbabebgbnbobrbscacscydadeeleneseteufafifofrglguhahawhehihrhthuhyidisitjajvkakkkmknkolalblnloltlvmgmimkmlmnmrmsmtmymrnenlnnnoocpaplpsptrorusasdsiskslsnsosqsrsusvswtatetgthtktltrttukuruzviyiyozhT) use_base_code)erAFAMARASAZBABEBGBNBOBRBSCACSCYDADEELENESETEUFAFIFOFRGLGUHAHAWHEHIHRHTHUHYIDISITJAJVKAKKKMKNKOLALBLNLOLTLVMGMIMKMLMNMRMSMTMY_MRNENLNNNOOCPAPLPSPTRORUSASDSISKSLSNSOSQSRSUSVSWTATETGTHTKTLTRTTUKURUZVIYIYOZHr )r% LANGUAGE_MAPs M/opt/pipecat/venv/lib/python3.12/site-packages/pipecat/services/gladia/stt.pylanguage_to_gladia_languager;sd Td Td  Td  T d  T d  T d  Td  Td  Td  Td  Td  Td  Td  Td  Td  T!d"  T#d$  T%d&  T'd(  T)d*  T+d,  T-d.  T/d0  T1d2  T3d4  T5d6  T7d8  T9d:  T;d<  e=d>  T?d@  TAdB  TCdD  TEdF  TGdH  TIdJ  TKdL  TMdN  TOdP  TQdR  TSdT  TUdV  TWdX  TYdZ  T[d\  T]d^  T_d`  Tadb  Tcdd  Tedf  Tgdh  Tidj  Tkdl  Tmdn  Todp  Tqdr  Tsdt  Tudv  Twdx  Tydz {d|  T}d~  Td@  TAdB  TCdD  TEdF  TGdH  TIdJ  TKdL  TMdN  TOdP  TQdR  TSdT  TUdV  TWdX  TYdZ  T[d\  T]d^  T_d`  Tadb  Tcdd  Tedf  Tgdh  Tidj  Tkdl  T T T T T T T T T T T T T TGdLL Hl$ GGceZdZdZddZy)_InputParamsDescriptorz?Descriptor for backward compatibility with deprecation warning.Nctj5tjdtjdtddddt S#1swYt SxYw)NalwayszGladiaSTTService.InputParams is deprecated and will be removed in a future version. Import and use GladiaInputParams directly instead. stacklevel)warningscatch_warnings simplefilterwarnDeprecationWarningr)selfobjobjtypes r__get__z_InputParamsDescriptor.__get__sT  $ $ &   ! !( + MME"   !  ! s 2AA#N)__name__ __module__ __qualname____doc__rrrrrs I !rrcleZdZUdZedZedzezed<edZ e e e fdzezed<edZ edzezed <ed Zedzezed <ed Zedzezed <edZedzezed<edZedzezed<edZedzezed<y)GladiaSTTSettingsa[Settings for GladiaSTTService. Parameters: language_config: Language detection and handling configuration. custom_metadata: Additional metadata to include with requests. endpointing: Silence duration in seconds to mark end of speech. maximum_duration_without_endpointing: Maximum utterance duration without silence. pre_processing: Audio pre-processing options. realtime_processing: Real-time processing features. messages_config: WebSocket message filtering options. enable_vad: Enable VAD to trigger end of utterance detection. ctSrrrrrzGladiaSTTSettings.W`r)default_factoryNlanguage_configctSrr rrrr zGladiaSTTSettings.rrcustom_metadatactSrr rrrr zGladiaSTTSettings.s)r endpointingctSrr rrrr zGladiaSTTSettings. r$maximum_duration_without_endpointingctSrr rrrr zGladiaSTTSettings.rrpre_processingctSrr rrrr zGladiaSTTSettings.rrrealtime_processingctSrr rrrr zGladiaSTTSettings.rrmessages_configctSrr rrrr zGladiaSTTSettings.s r enable_vad)rrrrrrrr__annotations__rdictstrrrfloatrintrrrrrrrboolrrrr r s :?O`9aO^d*Y6a9>O`9aOT#s(^d*Y6a,1BS,TK )TCH)D(#*y*@>C)>N'$.:HM)H1D89D:?O`9aO^d*Y6a*/@Q*RJt i'Rrr ceZdZUdZeZeed<eZddddddddddd de d d e d e d dzde de de de de ede e de e de ede dede ede effdZdZdefdZdede e fdZdee effd Zd!effd" Zd#edee efffd$ Zd!effd% Zd!effd& Zd'e de!e"dffd(Z#fd)Z$fd*Z%d+Z&d,Z'dee effd-Z(e) d:d.e d/ede e fd0Z*d1Z+d2Z,d'e fd3Z-d4Z.d5Z/d6Z0d7Z1d8e fd9Z2xZ3S);GladiaSTTServiceaSpeech-to-Text service using Gladia's API. This service connects to Gladia's WebSocket API for real-time transcription with support for multiple languages, custom vocabulary, and various processing options. Provides automatic reconnection, audio buffering, and comprehensive error handling. For complete API documentation, see: https://docs.gladia.io/api-reference/v2/live/init .. deprecated:: 0.0.62 Use :class:`~pipecat.services.gladia.config.GladiaInputParams` directly instead. _settingsNzhttps://api.gladia.io/v2/livewav/pcmi@T) regionurlencoding bit_depthchannels confidence sample_ratemodelparamsmax_buffer_sizeshould_interruptsettingsttfs_p99_latencyapi_keyr,)zus-westzeu-westr-r.r/r0r1r2r3r4r5r6r7r8c |rNtj5tjdtjdtddddt dddddddddd }| t d t d | |_| t d t | jNtj5tjdtjd tdddd| s| j | j}| j | j}| j | j}| j|_ | j|_ | j|_| j|_| j |_| j"|_| j$|_| j&r| j&|_n<| jr0|j)| j}|rt+|gd |_| |j-| t/|`d||dd|d|||_||_||_d|_||_||_||_d|_ d|_!d|_"tG|_$d|_%| |_&tOjP|_)d|_*| |_+y#1swYuxYw#1swYxYw)aInitialize the Gladia STT service. Args: api_key: Gladia API key for authentication. region: Region used to process audio. eu-west or us-west. Defaults to eu-west. url: Gladia API URL. Defaults to "https://api.gladia.io/v2/live". encoding: Audio encoding format. Defaults to ``"wav/pcm"``. bit_depth: Audio bit depth. Defaults to 16. channels: Number of audio channels. Defaults to 1. confidence: Minimum confidence threshold for transcriptions (0.0-1.0). .. deprecated:: 0.0.86 The 'confidence' parameter is deprecated and will be removed in a future version. No confidence threshold is applied. sample_rate: Audio sample rate in Hz. If None, uses service default. model: Model to use for transcription. .. deprecated:: 0.0.105 Use ``settings=GladiaSTTSettings(model=...)`` instead. params: Additional configuration parameters for Gladia service. .. deprecated:: 0.0.105 Use ``settings=GladiaSTTSettings(...)`` for runtime-updatable fields and direct init parameters for encoding/bit_depth/channels. max_buffer_size: Maximum size of audio buffer in bytes. Defaults to 20MB. should_interrupt: Determine whether the bot should be interrupted when Gladia VAD detects user speech. Defaults to True. settings: Runtime-updatable settings. When provided alongside deprecated parameters, ``settings`` values take precedence. ttfs_p99_latency: P99 latency from speech end to final transcript in seconds. Override for your deployment. See https://github.com/pipecat-ai/stt-benchmark **kwargs: Additional arguments passed to the STTService parent class. rzuThe 'confidence' parameter is deprecated and will be removed in a future version. No confidence threshold is applied.rrNz solaria-1F) r3r%rrrrrrrrr3r4znThe 'language' parameter is deprecated and will be removed in a future version. Use 'language_config' instead.) languagescode_switching)r2r8keepalive_timeoutkeepalive_intervalr7rr),rrrrrr rr3r%r.r/r0rrrrrrrrlanguage_to_service_languager apply_updatesuper__init___api_key_region_url _receive_task _encoding _bit_depth _channels _session_url _session_id_connection_active bytearray _audio_buffer _bytes_sent_max_buffer_sizeasyncioLock _buffer_lock _is_speaking_should_interrupt)rr9r,r-r.r/r0r1r2r3r4r5r6r7r8kwargsdefault_settings language_code __class__s rrDzGladiaSTTService.__init__sn ((* %%h/ :&  -  12 $      "7,=w G%*  "   "8-> ?*,,.))(3MMB*#$ ??.%H##/ & 0 0I??.%H393I3I 0/5/A/A ,??!E392G2G /7=7Q7Q 4393I3I 0.4.?.? +))7=7M7M$4__$($E$Efoo$VM$;I'4oe<(8    ) )( 3  #-  %        !"#!!"''[ /#LLN"!1O  @s2J=42K =K Kc:|jd|jdS)Nz [])namerMrs r__str__zGladiaSTTService.__str__s ))Bt//022rr&cy)zCheck if the service can generate performance metrics. Returns: True, indicating this service supports metrics generation. Trr_s rcan_generate_metricsz%GladiaSTTService.can_generate_metricss rr%ct|S)zConvert pipecat Language enum to Gladia's language code. Args: language: The Language enum value to convert. Returns: The Gladia language code string or None if not supported. )r)rr%s rrAz-GladiaSTTService.language_to_service_languages+844rc|j}|jxsd|jxsd|j|jxsd|j d}t |jxsi|d<t|dd<|j|j|d<|j|j|d<|jr|jjd |d <|jr|jjd |d <|jr|jjd |d <|jr|jjd |d<|S)Nr)r*r+)r.r/r2r0r3rpipecatrrT) exclude_nonerrrr)r(rIrJr2rKr3r!rpipecat_versionrrr model_dumprrr)rsr7s r_prepare_settingsz"GladiaSTTService._prepare_settingssc NN3).B+++!WW  '+1+<+<+B&C"#1@1B"#I. == $&'mmH] # 1 1 =66 ; <  *+*;*;*F*FTX*F*YH& '  )*)9)9)D)DRV)D)WH% & ./.C.C.N.N\`.N.aH* +  *+*;*;*F*FTX*F*YH& 'rframectKt||d{|jd{y77w)zStart the Gladia STT websocket connection. Args: frame: The start frame triggering service startup. N)rCstart_connectrrkr[s rrmzGladiaSTTService.starts3 gmE"""mmo #848688deltaclKt||d{}|s|S|j||S7w)zApply settings delta. Settings are stored but not applied to the active session. Args: delta: A settings delta. Returns: Dict mapping changed field names to their previous values. N)rC_update_settings _warn_unhandled_updated_settings)rrqchangedr[s rrsz!GladiaSTTService._update_settingss=077N --g68s 424cKt||d{|jd{|jd{y7577 w)zStop the Gladia STT websocket connection. Args: frame: The end frame triggering service shutdown. N)rCstop_send_stop_recording _disconnectros rrwzGladiaSTTService.stopsL gl5!!!'')))    ") s1AA AAAAAAActKt||d{|jd{y77w)zCancel the Gladia STT websocket connection. Args: frame: The cancel frame triggering service cancellation. N)rCcancelryros rr{zGladiaSTTService.cancels6 gnU###    $ rpaudioclK|jd{|j4d{|jj|t |j|j kDrot |j|j z }|j|d|_t d|j|z |_tj|d|ddddd{|jrL|jr@|jjtjur |j|d{dy747"7h#1d{7swYxxYw7'#t j"j$$r+}tj|d|d|_ Yd}~hd}~wwxYww)zRun speech-to-text on audio data. Args: audio: Raw audio bytes to transcribe. Yields: None (processing is handled asynchronously via WebSocket). Nrz) Audio buffer exceeded max size, trimmed z bytesz- Websocket closed while sending audio chunk: F)start_processing_metricsrUrPextendlenrRmaxrQr warningrN _websocketstater$OPEN _send_audio websockets exceptionsConnectionClosed)rr| trim_sizees rrun_sttzGladiaSTTService.run_stts++---$$ d d    % %e ,4%%&)>)>> 2 23d6K6KK %)%7%7 %C"#&q$*:*:Y*F#G $'PQZP[[abc d d  " "t4??;P;PTYT^T^;^ 0&&u---  ) . d d d d d.((99 0$'TUVTWXY*/'' 0sF4E F4E F4B-E F4'E(AF4+E)?E'E)F4 F4F4E$E E$ F4'E))F1!F,'F4,F11F4cK|jsa|j}|j|d{}|d|_|d|_t j |d|j|j d{t|!d{|jr=|js0|j|j|j|_ yyy77f7Rw)zxConnect to the Gladia service. Initializes the session if needed and establishes websocket connection. Nr-rLz Session URL: )rLrj _setup_gladiarMr info_connect_websocketrCrnrrH create_task_receive_task_handler _report_error)rr7responser[s rrnzGladiaSTTService._connect%s   --/H!//99H (D '~D  KK4&t/@/@.AB C%%'''g   ??4#5#5!%!1!1$2L2LTM_M_2`!aD $6?: ( s51C/C)AC/C+C/C-AC/+C/-C/cKt|d{d|_|jr*|j |jd{d|_|j d{y7Z7&7 w)zfDisconnect from the Gladia service. Cleans up tasks and closes websocket connection. NF)rCryrNrH cancel_task_disconnect_websocket)rr[s rryzGladiaSTTService._disconnect9sp g!###"'   ""4#5#56 6 6!%D ((*** $ 7 +s3A6A05A6 A2 A6*A4+A62A64A6cK |jr'|jjtjuryt j |dt |jd{|_d|_|j4d{d|_ dddd{|jdd{|jd{t j |dy77g7R#1d{7swYbxYw7P7:#t$r%}|jd||d{7d}~wwxYww) z-Establish the websocket connection to Gladia.NzConnecting to Gladia WebSocketTr on_connectedz Connected to Gladia WebSocketzUnable to connect to Gladia:  error_msg exception)rrr$rr debugwebsocket_connectrLrNrUrQ_call_event_handler_send_buffered_audio Exception push_errorrrs rrz#GladiaSTTService._connect_websocketHs! 4??#8#8EJJ#F LLD6!?@ A$5d6G6G$HHDO&*D #(( % %#$  % %**>: : :++- - - LLD6!?@ AI % % % % % ; . //.KA3,O[\/] ] ]  sD>2D D>0D 'C.( D C0 D C4 D C2 D 8D 9D D D -D>.D 0D 2D 4D:C= ;DD D D;D6/D20D66D;;D>cK |jr`|jjtjur:t j |d|jj d{d|_|jdd{y7%#t$r)}|jd||d{7Yd}~Pd}~wwxYw7<#d|_|jdd{7wxYww)z)Close the websocket connection to Gladia.z$ Disconnecting from Gladia WebSocketNzError closing websocket: ron_disconnected) rrr$rr rcloserrrrs rrz&GladiaSTTService._disconnect_websocketas >4??#8#8EJJ#F v%IJKoo++---#DO**+<= = = . Z//.Gs,KWX/Y Y Y Z >#DO**+<= = =swC0A'B+B,B0C0 C C0B CC5B86C;C CC C0 C-&C)'C--C0c Ktj4d{}i}|jr|j|d<|j|jd|j i||4d{}|j r:|jd{cdddd{cdddd{S|jd{}tj|d|jd|xs |jt|d|jd|77777x7a#1d{7swYnxYwdddd{7y#1d{7swYyxYww)Nr,z X-Gladia-Key)headersjsonr4z Gladia error: z: z& Failed to initialize Gladia session: z - )aiohttp ClientSessionrFpostrGrEokrtextr errorstatusreasonr)rr7sessionr4r error_texts rrzGladiaSTTService._setup_gladiams]((*  gF||#'<<x || '7 $   ;;!)0       (0}}!6JLL&/@:C`QYQ`Q`Bab$& FxFWWZ[eZfg#  1   "7         sE*D!E*A E,D$-E0 D.D&D. E D(!E% E*1D*2E*7D. D, AD.!E*$E&D.(E*E*,D..E 4D75E <E E*EE*E'E E'#E* transcriptis_finalc@K|jd{y7wr)stop_processing_metrics)rrrr%s r_handle_transcriptionz&GladiaSTTService._handle_transcriptions**,,,s cK|jjr |jrytj|dd|_|j t d{|jr|jd{yy7*7w)zHandle speech start event from Gladia. Broadcasts UserStartedSpeakingFrame and optionally triggers interruption when VAD is enabled. Nz User started speakingT) r(rrVr rbroadcast_framerrWbroadcast_interruptionr_s r_on_speech_startedz#GladiaSTTService._on_speech_startedsz ~~((D,=,=  v345 ""#;<<<  ! !--/ / / " = /s$AB B#B B B  B cK|jjr |jsyd|_|jtd{t j |dy7w)zoHandle speech end event from Gladia. Broadcasts UserStoppedSpeakingFrame when VAD is enabled. NFz User stopped speaking)r(rrVrrr rr_s r_on_speech_endedz!GladiaSTTService._on_speech_endedsV ~~((0A0A !""#;<<< v345 =sAA%A#A%c:K|jr|jjtjurbt j |j d}dd|id}|jjtj|d{yyy7w)z,Send audio chunk with proper message format.zutf-8 audio_chunkchunk)typedataN) rrr$rbase64 b64encodedecodesendrdumps)rr|rmessages rrzGladiaSTTService._send_audiosy ??t44 B##E*11':D,woFG//&&tzz'':; ; ; C? B* B B*BB*B'B B'#B*cK|jr`|jjtjur9|jj t j ddid{yyy7w)Nrstop_recording)rrr$rrrrr_s rrxz%GladiaSTTService._send_stop_recordingsT ??t44 B//&&tzz6;K2L'MN N N C? NsA&A2(A0)A2cH|jr |jStd)zGet the current WebSocket connection. Returns: The WebSocket connection. Raises: Exception: If WebSocket is not connected. zWebsocket not connected)rrr_s r_get_websocketzGladiaSTTService._get_websockets! ???? "122rc K|j23d{} tj|}|ddk(r~|jdrm|dd}|j4d{|d}||j kDr*||j z }|j |d|_||_dddd{nJ|ddk(r|dd }|d }|d }|dd } | rV|jt||jt|| d{|j|| |d{n|jt||jt|| d{n|ddk(rM|dd} |dd} | d } | d } | | k7ro|jt| dt| d{nA|ddk(r|jd{n |ddk(r|jd{777Y#1d{7swYxYw7777m7N7/#tj $rt#j$|d|YLwxYw6yw)zxReceive and process websocket messages. Continuously processes messages from the websocket connection. Nrr acknowledgedr byte_ranger+r utterancer%rr)result)rrr% translationtranslated_utteranceoriginal_language speech_start speech_endz Received non-JSON message: )rrloadsgetrUrQrP push_framer_user_idr!rrrrrJSONDecodeErrorr r)rrcontentrend_byterrr%rrrrtranslated_languagers r_receive_messagesz"GladiaSTTService._receive_messagess "002= O= O'< O**W-6?m3 N8S!(!>J#0088#-a=#d&6&66(043C3C(CI151C1CIJ1OD./7D, 888V_ 4 ' z*J'"6v">K*.??"oo, +R1A1CEX V_611333V_ 4//111w= O88888  41'' O$'CG9MN Oy3sI%I#HI#I%AH/#H$H/'?H & H/1H 2AH/H"H/,H%-9H/&H''AH/;H)< H/H+ H/=H->H/I%I#H/ H/ H HH H/%H/'H/)H/+H/-H//-I I%I  I%silencecBK|jdd{y7w)zSend an empty audio chunk to keep the Gladia connection alive. Args: silence: Silent PCM audio bytes (unused, Gladia accepts empty chunks). rN)r)rrs r_send_keepalivez GladiaSTTService._send_keepalives s###s r)4rrrrr Settingsr r InputParamsrr"rr$rr#rr%rDr`rbrrAr!rrjrrmrsr rwr r{rrr rrnryrrrr"rrrrrrxrrr __classcell__)r[s@rr'r's@ !H  )*K 8<2!&*%)#.2/!%04,;!_2_2,-4 _2  _2  _2_2_2UO_2c]_2}_2*+_2_2_2,-_2 #5/!_2B3d 5X 5(3- 5'4S>'R,=$sCx.6!!!+!5^E4K-H>b( +2 >DcN,IM--)--9A#-- 0 6rs  (99.   `_8<F/?,F)oH(oHx}oHf ! ! S S S